Kamil Braun

> In CQL, schema-changing operations like CREATE KEYSPACE return a response to the client telling it "I did a schema change, please wait for the schema to settle" - and...

@soyacz you didn't mark that this issue is a regression or not. Why? Is this a new test that you weren't running before?

> So, having c-s from 2022.2.dev-0.20220330.eef4cbb20a51 does the driver does it? Maybe it's time to update our loaders? The driver did it since forever. It shouldn't have changed across versions.

> It's not a new test, but possibly we didn't run to such race condition in the past so I was skeptical to mark it as regression. If it never...

So it looks like this is not a regression, the problem was there since forever, the test at some point got modified and started catching it: https://github.com/scylladb/scylladb/issues/17124#issuecomment-1960836398 @fruch mentioned: >...

Test was disabled in gossip mode, should not be flaky in raft-topo mode --- closing. We can reopen if we observe in raft-topo.

Also nit: commit messages should be prefixed with the name of the module they're modifying, for example `test/pylib: Introduce Start-Stop Lock` (we have no concrete rule what exactly the prefix...

Changed Fixes markers in commit message to: ``` Fixes: scylladb/scylladb#16472 Fixes: scylladb/scylladb#16651 ``` (so bot can correctly react)

Due to the SM backup issue, marking this as release blocker. (@mykaul please review the decision.)

Taking off release blocker label then.